BNY uses Azure OpenAI in Foundry Models to scale trust in accounting operations

As transaction volumes, product complexity, and regulatory expectations grow, financial institutions face increasing pressure to scale investigative and analytical expertise without proportionally increasing cost, risk, or overhead.

With Azure OpenAI in Foundry Models, Microsoft Foundry, and the Microsoft co-engineering approach, BNY is building an AI research capability to help clients investigate complex data issues, evaluate context, and support operational decision-making.

The initiative is designed to help the BNY Data & Analytics team explore a new financial services operating model that embeds intelligence into operations while maintaining the accuracy, auditability, and control required in regulated environments.

Scaling trust in a more complex financial world

The financial services industry runs on trust. Every report, decision, and statement depends on processes designed to ensure accuracy, consistency, and compliance. As transaction volumes grow and financial products become more sophisticated, maintaining trust becomes increasingly difficult.

Investigating data discrepancies sits at the center of this challenge. When differences emerge across systems, records, and financial positions, fund accounting teams must research the root cause, evaluate context across multiple sources, and determine the right course of action, all while maintaining the accuracy and auditability required by regulators and clients. That process has long relied on significant human expertise.

“In our conversations with clients, we consistently hear that the challenge is not simply processing more transactions. What matters is their ability to maintain the same level of trust, accuracy, and control as complexity and volume increases.”

Tony Evans, Global Head of Relationship Management and Sales, Data & Analytics, BNY

For BNY, the opportunity was more than automating a workflow. It was exploring whether AI could help scale trust, consistency, and operational performance in a way traditional automation could not. “In our conversations with clients, we consistently hear that the challenge is not simply processing more transactions. What matters is their ability to maintain the same level of trust, accuracy, and control as complexity and volume increases,” says Tony Evans, Global Head of Relationship Management and Sales, Data & Analytics, BNY.

Moving beyond automation

Traditional automation can execute predefined rules and streamline repetitive tasks. But many investigative scenarios still require human judgment to understand context, evaluate causes, and determine the next course of action.

A collaboration grew from BNY’s participation in the Microsoft CEO Co-Innovation program, where the organizations explored new opportunities to apply AI to longstanding financial services challenges. The initiative emerged as one of the concepts selected for joint exploration and development.

“We weren’t looking to build another rules-based automation tool. We were exploring whether AI could become an operational participant in the accounting and investment operations process and Foundry fit the bill.”

Jay Hinton, Head of Global Product, Data & Analytics, BNY

BNY worked with Microsoft to explore a different approach. Using Microsoft Azure OpenAI in Foundry Models, Microsoft Foundry, and the Microsoft co-engineering model, the organizations developed an AI research capability embedded directly within BNY’s Data & Analytics Accounting platform. The solution is designed to investigate complex data issues, analyze context across multiple sources, identify likely root causes, and support resolution workflows within existing operational processes. For BNY, the Microsoft role extends beyond providing AI models.

Azure OpenAI delivers advanced reasoning capabilities, while Foundry provides the evaluation, governance, observability, and operational controls needed to support AI in regulated financial environments. “We weren’t looking to build another rules-based automation tool. We were exploring whether AI could become an operational participant in the accounting and investment operations process and Foundry fit the bill,” says Jay Hinton, Head of Global Product, Data & Analytics, BNY.

From human investigation to AI-assisted reasoning

One of the most important aspects of the initiative is how the system approaches data discrepancies. In the past, operational teams were responsible for gathering information, evaluating potential causes, and determining how exceptions should be resolved.

“The opportunity is not just to automate work, but to also deepen collaboration through our partner ecosystem and embed Azure intelligence directly into the processes that drive operational performance.”

Deanna Lanier, Global Head of Strategy, Data & Analytics, BNY

The solution is designed to participate in that process by classifying issues, analyzing relevant context across ledger and subledger sources, identifying likely root causes, and recommending next steps. Beyond faster processing, it demonstrates how AI can support workflows that have always depended on human investigation.

For BNY, this became a proving ground for embedding intelligence directly into financial operations. “The opportunity is not just to automate work, but to also deepen collaboration through our partner ecosystem and embed Azure intelligence directly into the processes that drive operational performance,” says Deanna Lanier, Global Head of Strategy, Data & Analytics, BNY.

Establishing a foundation for AI-driven financial operations

The solution is currently entering a pilot phase, with goals that include reducing manual research effort while improving operational consistency and efficiency. More importantly, the initiative is helping BNY explore how AI can be applied to complex operational processes where scale, accuracy, governance, and trust must coexist.

By combining BNY’s operational expertise with Azure OpenAI and Foundry, the organization is helping define what production-grade AI can look like in financial services. The long-term opportunity extends beyond automation. It’s creating a new operating model where intelligence helps financial institutions scale trust without proportionally increasing operational overhead.

Discover more about BNY Data & Analytics on LinkedIn.